Вы находитесь на странице: 1из 23

To appear in the ASCE Journal of Engineering Mechanics

Phenomenological Model of a Magnetorheological Damper


B.F. Spencer Jr.,1 S.J. Dyke,2 M.K. Sain3 and J.D. Carlson4

Abstract
Semi-active control devices have received significant attention in recent years because they
offer the adaptability of active control devices without requiring the associated large power sources. Magnetorheological (MR) dampers are semi-active control devices that use MR fluids to produce controllable dampers. They potentially offer highly reliable operation and can be viewed as
fail-safe in that they become passive dampers should the control hardware malfunction. To develop control algorithms that take maximum advantage of the unique features of the MR damper,
models must be developed that can adequately characterize the dampers intrinsic nonlinear behavior. Following a review of several idealized mechanical models for controllable fluid dampers,
a new model is proposed that can effectively portray the behavior of a typical magnetorheological
damper. Comparison with experimental results for a prototype damper indicates that the model is
accurate over a wide range of operating conditions and is adequate for control design and analysis.

Introduction
Passive and active control systems represent the two ends of the spectrum in the use of supplemental damping strategies for response reduction in civil engineering structures subjected to
strong earthquakes and severe winds (see, for example, Soong 1990; Soong, et al., 1991; Housner
and Masri 1990, 1993; Housner, et al., 1994). On the other hand, semi-active control systems
combine the best features of both approaches, offering the reliability of passive devices, yet maintaining the versatility and adaptability of fully active systems. According to presently accepted
definitions, a semi-active control device is one that has properties that can be adjusted in real time
but cannot input energy into the system being controlled. Such devices typically have very low
power requirements, which is particularly critical during seismic events when the main power
source to the structure may fail. Moreover, because many active control systems for civil engineering applications operate primarily to modify structural damping, preliminary studies indicate
that semi-active control strategies can potentially achieve the majority of the performance of fully
active systems. Various semi-active devices have been proposed which utilize forces generated by
surface friction or viscous fluids to dissipate vibratory energy in a structural system. Examples of
such devices that have been considered for civil engineering applications include variable orifice
dampers (e.g., Shinozuka, et al. 1992; Kawashima, et al. 1992; Mizuno, et al. 1992; Constanti-

1. Professor, Department of Civil Engineering and Geological Sciences, University of Notre Dame, Notre Dame, IN
46556, Member ASCE.
2. Doctoral Candidate and Graduate Research Assistant, Department of Civil Engineering and Geological Sciences,
University of Notre Dame, Notre Dame, IN 46556, Student Member ASCE.
3. Freimann Professor, Department of Electrical Engineering, University of Notre Dame, Notre Dame, IN 46556,
Member ASCE.
4. Engineering Fellow, Lord Corporation, Mechanical Products Division, Thomas Lord Research Center, 405 Gregson Drive, Cary, NC 27511-7900.
Spencer, Dyke, Sain & Carlson

March 10, 1996

nou, et al. 1993; Constantinou and Symans 1994; Sack and Patten 1994; Sack et al. 1994; Patten,
et al. 1994, Kurata, et al. 1994), friction controllable braces (e.g., Akbay and Aktan 1990 & 1991;
Dowdell and Cherry 1994; Cherry 1994), friction controllable isolators (e.g., Feng and Shinozuka
1990; Kawashima, et al. 1992) and variable stiffness devices (Kobori, et al. 1993; Inaudi and
Kelly 1994).
Another class of semi-active devices uses controllable fluids. The essential characteristic of
controllable fluids is their ability to reversibly change from a free-flowing, linear viscous fluid to a
semi-solid with a controllable yield strength in milliseconds when exposed to an electric or magnetic field. Two fluids that are viable contenders for development of controllable dampers are: (i)
electrorheological (ER) fluids and (ii) magnetorheological (MR) fluids. Although the discovery of
both ER and MR fluids dates back to the late 1940s (Winslow 1947; Winslow 1949; Rabinow
1948), research programs have primarily concentrated on ER fluids. A number of researchers
have considered electrorheological fluid dampers for civil engineering applications (e.g., Ehrgott
and Masri 1992, 1994; Gavin and Hanson 1994; Gavin, et al., 1994; Gavin, 1994; Makris, et al.,
1995).
Recently developed MR fluids appear to be an attractive alternative to ER fluids for use in
controllable fluid dampers (Carlson 1994; Carlson and Weiss 1994; Carlson, et al. 1995; see also:
http://www.rheonetic.com/mrfluid/). MR fluids are the magnetic analogs of electrorheological fluids and typically consist of micron-sized, magnetically polarizable particles dispersed in a carrier
medium such as mineral or silicone oil. When a magnetic field is applied to the fluids, particle
chains form, and the fluid becomes a semi-solid and exhibits viscoplastic behavior similar to that
of ER fluids. Transition to rheological equilibrium can be achieved in a few milliseconds, allowing construction of devices with high bandwidth. Additionally, Carlson and Weiss (1994) indicated that the achievable yield stress of an MR fluid is an order of magnitude greater than its ER
counterpart and that MR fluids can operate at temperatures from 40 to 150oC with only slight
variations in the yield stress. Moreover, MR fluids are not sensitive to impurities such as are commonly encountered during manufacturing and usage, and little particle/carrier fluid separation
takes place in MR fluids under common flow conditions. Further a wider choice of additives (surfactants, dispersants, friction modifiers, anti-wear agents, etc.) can generally be used with MR fluids to enhance stability, seal life, bearing life, etc., since electro-chemistry does not affect the
magneto-polarization mechanism. The MR fluid can be readily controlled with a low voltage
(e.g., ~1224V), current-driven power supply outputting only ~12 amps. Recognizing the significant potential of devices based on MR fluids, a number of researchers have recently undertaken
their study (see, for example, Bossis and Lemaire 1991; Carlson 1994; Carlson and Weiss 1994;
Demchuk 1993; Dyke, et al. 1996a,b; Fedorov 1992; Grasselli, et al. 1993; Horvath and Kopcansky 1993; Kabakov and Pabat 1990; Kashevskii 1990; Kordonsky 1993a,b; Kordonsky, et al.
1990, 1993; Lemaire, et al. 1994; Minagawa, et al. 1994; Pabat 1990; Savostyanov 1992; and
Shulman, et al. 1986, 1989; Spencer, et al. 1996a,b).
To evaluate the usefulness of MR devices in response reduction for civil engineering structures, a prototype MR damper has been obtained from the Lord Corporation of Cary, North Carolina for laboratory testing. The first step in this evaluation is to develop a high fidelity model for
use in control design and analysis. This task is challenging because the MR damper, like most
semi-active control means, is a highly nonlinear device. The experimental setup used to obtain the
data necessary for identification of a model for the prototype MR damper is discussed in the next
section. Then, following a brief review of several idealized models of controllable fluid dampers,
Spencer, Dyke, Sain & Carlson

March 10, 1996

a new phenomenological model is proposed that is numerically tractable and effectively portrays
the behavior of an MR damper. Based on the experimental data, a least-squares optimization
method is employed to determine appropriate parameters for the analytical model. Comparison
with the data indicates that the model presented is accurate for a wide range of operating conditions and is appropriate for use in control algorithm development and system evaluation.

Experimental Setup and Sinusoidal Response Data


To take full advantage of the unique features of the MR damper in control applications, a
model must be developed that can accurately reproduce the behavior of the MR damper. The load
frame shown in Fig. 1 was designed and built for the purpose of obtaining the MR damper response data necessary for identification studies. In this setup, a double-ended hydraulic actuator,
manufactured by Nopak, was employed to drive the damper. The actuator had a 3.8 cm diameter
cylinder and a 30.5 cm stroke and was fitted with low-friction Teflon seals to reduce nonlinear effects. A servo-valve, made by Dynamic Valves, Inc., with a nominal operational frequency range
of 045 Hz was used to control the actuator. A Schaevitz linear variable differential transformer
(LVDT) was used to measure the displacement of the piston-rod of the MR damper, and an Omega load cell with a range of 4540 N was included in series with the damper to measure the output force. The data acquisition system employed consisted of eight Syminex XFM82 3-decade
programmable antialiasing filters having a 90 dB signal to noise ratio and simultaneous sample
and hold, an Analogic LSDAS-16-AC-mod2 data acquisition board having 16 bit A/D converters,
an Analogic CTRTM-05 counter-timer board, a Gateway 2000 P5-90 computer, and the HEM
Data Corporation Snap-Master software. Using this experimental setup, dynamic responses of the
damper can be measured for a wide range of prescribed displacement wave forms, including sine,
step, triangle and pseudo random.
The prototype MR damper obtained for evaluation is a fixed orifice damper filled with a magnetorheological fluid. The MR fluid is a proprietary formulation, VersaFloTM MRX-135GD developed by the Lord Corporation, that consists of micron-size, magnetically-soft iron randomly
dispersed in a hydrocarbon oil along with additives that promote homogeneity and inhibit gravitational settling. It has a density of 3.28 g/cm3. The damper is 21.5 cm long in its extended position,
and the main cylinder is 3.8 cm in diameter. The main cylinder houses the piston, the magnetic
circuit, an accumulator and 50 ml of MR fluid, and the damper has a 2.5 cm stroke. As shown in
Fig. 2, the MR fluid valve is contained within the damper piston and consists of an annular flow
channel having an inner diameter of 27 mm and an outer diameter of 28 mm. The magnetic field
is applied radially across the resulting 0.5 mm dimension, perpendicular to the direction of fluid
flow. The total axial length of the flow channel is 15 mm of which 7 mm is exposed to the applied
magnetic field. Thus, the total volume of fluid that sees the magnetic field at any instant is about

Hydraulic Actuator

MR Damper

Force
Transducer

Figure 1. Test Setup for MR Damper Identification.


Spencer, Dyke, Sain & Carlson

March 10, 1996

Wires to
Electromagnet
MR Fluid

Magnetic
Choke

Accumulator

Rod

Piston

Figure 2. Schematic of MR Damper.

0.3 ml. The magnetic field can be varied from 0 to 200kA/m for currents of 0 to 1 amp in the electromagnet coil, which has a resistance of 4 . The total inductance of the MR fluid valve is 40
mH, resulting in an L/R time constant of 10 msec. The peak power required is less than 10 watts,
which would allow the damper to be operated continuously for more than an hour on a small camera battery. For this example, the current for the electromagnet is supplied by a linear current driver running off of 120 volts AC and generates a 0 to 1 amp current that is proportional to a
commanded DC input voltage in the range 03 V. With this power supply, the damper typically
reaches rheological equilibrium in less than 6.5 msec after applying the magnetic field (Carlson
and Weiss 1994). Forces of up to 3000 N can be generated with the device, and are stable over a
broad temperature range, varying less than 10% in the range of 40 to 150 degrees Celsius.
Using the setup depicted in Fig. 1, a series of preliminary tests was conducted to measure the
response of the damper under various loading conditions. In each test, the hydraulic actuator was
driven with a sinusoidal signal with a fixed frequency, and the voltage applied to the prototype
MR damper was held at a constant level. A wide range of frequencies, amplitudes and voltage
levels were considered. The data was sampled at 256 Hz. The velocity response was calculated
from the measured displacements via a central difference approximation.
The response of the MR damper due to a 2.5 Hz sinusoid with an amplitude of 1.5 cm is
shown in Fig. 3 for four constant voltage levels, 0 V, 0.75 V, 1.5 V, and 2.25 V, being applied to
the power amplifier for the device. These voltages correspond to 0 A, 0.25 A, 0.5 A and 0.75 A,
respectively. The force generated as a function of time is shown in Fig. 3a, the force-displacement
loop is shown in Fig. 3b and the force-velocity loop is shown in Fig. 3c. Note that the force-displacement loops in Fig. 3b progress along a clockwise path with increasing time, whereas the
force-velocity loops in Fig. 3c progress along a counter-clockwise path with increasing time.
In Fig. 3, the effects of changing the magnetic field are readily observed. At 0 V the MR
damper primarily exhibits the characteristics of a purely viscous device (i.e., the force-displacement relationship is approximately elliptical, and the force-velocity relationship is nearly linear).
However, as the voltage increases, the force required to yield the MR fluid in the damper increases and produces behavior associated with a plastic material in parallel with a viscous damper, i.e.,
Bingham plastic behavior (Shames and Cozzarelli 1992). Also, notice that the increase in force

Spencer, Dyke, Sain & Carlson

March 10, 1996

1500
0V
0.75 V
1.5 V
2.25 V

Force (N)

1000
500
0
500
1000
1500
0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

Time (sec)

Force (N)

a) Force vs. Time


1500

1500

1000

1000

500

500

500

500

1000

1000

1500
2

1500
1

Displacement (cm)

b) Force vs. Displacement

10

10

Velocity (cm/sec)

c) Force vs. Velocity

Figure 3. Experimentally Measured Force for 2.5 Hz Sinusoidal


Excitation with an Amplitude of 1.5 cm.

for a given increase in the applied voltage is approximately linear for voltages between 02.25 V.
For the particular damper tested, saturation of the MR effect occurs above 2.25 V.
From Fig. 3, observe that the force produced by the damper is not centered at zero. This effect is due to the presence of an accumulator in the MR damper, which consists of a bladder within the main cylinder (see Fig. 2) that is filled with nitrogen pressurized at 300 psi. The
accumulator helps prevent cavitation in the fluid during normal operation and accounts for the
volume of fluid displaced by the piston rod as well as thermal expansion of the fluid. From a phenomenological perspective, the accumulator acts like a spring in the damper. In the experimental
data provided in Fig. 3, the presence of the accumulator produces an offset in the measured damper force and a slight vertical widening of the response loops in the force-velocity plot. To obtain
an effective model of the MR damper, the stiffness associated with the accumulator must be taken
into account.
Another interesting feature of the data that is important to note is seen in the force-velocity
responses shown in Fig. 3c. Focusing attention on the upper branch of the force-velocity curve,
which corresponds to decreasing velocities (i.e., negative accelerations, and therefore positive positions), for large positive velocities, the force in the damper varies linearly with velocity. However, as the velocity decreases and before it becomes negative, the force-velocity relationship is no
longer linear, decreasing rapidly and smoothly. This roll-off in the force at small velocities is due
to bleed or blow-by of fluid between the piston and the cylinder and is necessary to eliminate
harshness from the subjective feel of the damper in vehicular applications. This type of behavior
will be sought in a model of the MR damper.

Spencer, Dyke, Sain & Carlson

March 10, 1996

Both nonparametric and parametric models have been considered to model the observed behavior of controllable fluid dampers. Ehrgott and Masri (1994) presented a nonparametric approach for modeling ER fluid dampers by assuming that the damper force could be written in
terms of Chebychev polynomials in the damper velocity and acceleration. McClamroch and
Gavin (1995) followed a similar approach in modeling an ER device. One of the difficulties in this
approach is that the resulting models are often quite complex. Using basic mechanics, Kamath
and Wereley (1996) and Makris, et al. (1996) have developed parametric models to characterize
ER fluids and fluid devices. Alternatively, parametric models based on simple mechanical idealizations have been considered by Stanway, et al. (1985, 1987) and Gamota and Filisko (1991) to
describe the behavior of controllable fluids and fluid dampers. Such an approach is advocated
herein. The next section examines the effectiveness of several idealized mechanical models for
predicting the response of the prototype MR damper, and a new model is proposed that addresses
a number of shortcomings associated with these models.

Mechanical Model Formulation


The stress-strain behavior of the Bingham viscoplastic model (Shames and Cozzarelli, 1992)
is often used to describe the behavior of MR (and ER) fluids. In this model, the plastic viscosity is
defined as the slope of the measured shear stress versus shear strain rate data. Thus, for positive
values of the shear rate, , the total stress is given by
= y ( field ) +

(1)

where y ( field ) is the yield stress induced by the magnetic (or electric) field and is the viscosity
of the fluid.
Based on this model of the rheological behavior of ER fluids, Stanway, et al. (1985, 1987)
proposed an idealized mechanical model, denoted the Bingham model, for the behavior of an ER
damper. The Bingham model consists of a Coulomb friction element placed in parallel with a viscous damper, as shown in Fig. 4. In this model, for nonzero piston velocities, x , the force generated by the device given by
F = f c sgn ( x) + c 0 x + f 0

(2)

where c 0 is the damping coefficient and f c is the frictional force, which is related to the fluid
yield stress. An offset in the force f 0 is included to account for the nonzero mean observed in the

c0

F f0

Figure 4. Bingham Model of a Controllable Fluid


Damper (Stanway, et al. 1985, 1987).
Spencer, Dyke, Sain & Carlson

March 10, 1996

measured force due to the presence of the accumulator. Note that if at any point the velocity of the
piston is zero, the force generated in the frictional element is equal to the applied force.
To assess its ability to predict the behavior of the MR damper, the model in (2) was fit to the
2.5 Hz sinusoidal response data shown in Fig. 3 for the case in which the command voltage to the
current driver was a constant 1.5 V. The parameters chosen are f c = 670 N, c 0 = 50 N sec/cm
and f 0 = 95 N. Fig. 5 shows a comparison between the predicted and experimentally obtained
responses. Although the force-displacement behavior appears to be reasonably modeled, examination of the force-velocity response and the temporal variation of the force shows that the behavior of the damper is not captured, especially for velocities that are near zero. In particular, this
model does not exhibit the nonlinear force-velocity response observed in the data for the case
when the acceleration and velocity have opposite signs (or alternatively, when the velocity and the
displacement have the same sign) and the magnitude of the velocities are small. While this model
may be adequate for response analysis, it is not adequate for control analysis. Notice that the model predicts a one-to-one relationship between the force and velocity, but the experimentally obtained data is not one-to-one. Furthermore, at zero velocity, the measured force has a positive
value when the acceleration is negative (positive displacements), and a negative value when the
acceleration is positive (negative displacement). This behavior must be captured in a mathematical model to adequately characterize the device for control applications.
Also focusing on predicting the behavior of ER materials, Gamota and Filisko (1991) proposed an extension of the Bingham model, which is given by the viscoelastic-plastic model
shown in Fig. 6. The model consists of the Bingham model (i.e., a frictional element in parallel
with a dashpot) in series with a standard model of a linear solid (Shames and Cozzarelli, 1992).
The governing equations for this model are given by

1500

Force (N)

Experimental
Predicted

1500
0

0.2

0.4

0.6

0.8

12

Time (sec)

a) Force vs. Time


1500

Force (N)

1500

1500
2

Displacement (cm)

b) Force vs. Displacement

1500
12

Velocity (cm/sec)

c) Force vs. Velocity

Figure 5. Comparison Between the Predicted and Experimentally


Obtained Responses for the Bingham Model.
Spencer, Dyke, Sain & Carlson

March 10, 1996

c0

c1

x1

x3

x2
k2

k1

F f0

Figure 6. Model proposed by Gamota and Filisko (1991).

F = k 1 ( x 2 x 1 ) + c 1 ( x2 x1 ) + f 0

= c 0 x1 + f c sgn ( x1 ) + f 0
,

= k2 ( x3 x2) + f0

F > fc

(3)

F = k 1 ( x 2 x 1 ) + c 1 x2 + f 0
,
= k2 ( x3 x2) + f0

F fc

(4)

where c 0 is the damping coefficient associated with the Bingham model and k 1 , k 2 and c 1 are
associated with the linear solid material. Note that when F f c , x1 = 0 .
Again, parameters for the model in (3), (4) were determined to fit the 2.5 Hz data shown in
Fig. 3 for the case where the voltage to the current driver was 1.5 V. The parameters chosen are
4
6
f c = 670 N, c 0 = 5000 N sec/cm, c 1 = 1300 N sec/cm, k 1 = 5 10 N/cm, k 2 = 2 10
N/cm and f 0 = 95 N. A comparison between the predicted responses and the corresponding experimental data is provided in Fig. 7. As might be expected, this model can portray the force-displacement behavior of the damper well. In addition, it possesses force-velocity behavior that more
closely resembles the experimental data. However, the governing equations (3), (4) are extremely
stiff, making them difficult to deal with numerically. Numerical integration of (3), (4) for the parameters given previously required a time step on the order of 10 6 sec. Note that a decrease in
the damping, c 1 , can produce the nonlinear roll-off observed in the experimental force-velocity
relationship as the velocity approaches zero, but then even smaller time steps are required to simulate the system. The numerical challenges of this model constitute its main shortcoming, which
was also noted in Ehrgott and Masri (1994).
One model that is numerically tractable and has been used extensively for modeling hysteretic systems is the Bouc-Wen model (Wen 1976). The Bouc-Wen model is extremely versatile and
can exhibit a wide variety of hysteretic behavior. A schematic of this model is shown in Fig. 8.
The force in this system is given by
F = c 0 x + k 0 ( x x 0 ) + z

(5)

where the evolutionary variable z is governed by


z = x z z

Spencer, Dyke, Sain & Carlson

n1

x z + Ax .

March 10, 1996

(6)

1500

Force (N)

Experimental
Predicted

1500
0

0.2

0.4

0.6

0.8

Time (sec)

Force (N)

a) Force vs. Time


1500

1500

1500
2

Displacement (cm)

b) Force vs. Displacement

1500
12

12

Velocity (cm/sec)

c) Force vs. Velocity

Figure 7. Comparison Between the Predicted and Experimentally


Obtained Responses for the Gamota and Filisko Model.

x
Bouc-Wen

k0

c0

Figure 8. Bouc-Wen Model of the MR Damper.

By adjusting the parameters of the model , and A , one can control the linearity in the unloading and the smoothness of the transition from the pre-yield to the post-yield region. In addition,
the force f 0 due to the accumulator can be directly incorporated into this model as an initial deflection x 0 of the linear spring k 0 .
A set of parameters was determined to fit the response of the Bouc-Wen model to the experimentally measured response of the MR damper shown in Fig. 3 (2.5 Hz sinusoidal displacement
and a constant applied voltage of 1.5V). The parameters for the model in (5), (6) were chosen to
be = 880 N/cm, c 0 = 50 N sec/cm, k 0 = 25 N/cm, = 100 cm-2, = 100 cm-2,
n = 2 , A = 120 and x 0 = 3.8 cm. A comparison between the predicted responses and the corSpencer, Dyke, Sain & Carlson

March 10, 1996

1500

Force (N)

Experimental
Predicted

1500
0

0.2

0.4

0.6

0.8

12

Time (sec)

Force (N)

a) Force vs. Time


1500

1500

1500
2

Displacement (cm)

b) Force vs. Displacement

1500
12

Velocity (cm/sec)

c) Force vs. Velocity

Figure 9. Comparison Between the Predicted and Experimentally


Obtained Responses for the Bouc-Wen Model.

responding experimental data is provided in Fig. 9. The Bouc-Wen model predicts the force-displacement behavior of the damper well, and it possesses force-velocity behavior that more closely
resembles the experimental data. However, similar to the Bingham model, the nonlinear force-velocity response of the Bouc-Wen model does not roll-off in the region where the acceleration and
velocity have opposite signs and the magnitude of the velocities are small. To better predict the
damper response in this region, a modified version of the system in Fig. 8 is proposed, as shown
in Fig. 10. To obtain the governing equations for this model, consider only the upper section of
the model. The forces on either side of the rigid bar are equivalent; therefore,
c 1 y = z + k 0 ( x y ) + c 0 ( x y)

(7)

where the evolutionary variable z is governed by


z = x y z z

n1

( x y) z + A ( x y) .

(8)

Solving (7) for y results in


1
y = ----------------------- { z + c 0 x + k 0 ( x y ) } .
( c0 + c1)

(9)

The total force generated by the system is then found by summing the forces in the upper and lower sections of the system in Fig. 10, yielding

Spencer, Dyke, Sain & Carlson

March 10, 1996

10

Bouc-Wen
c1

k0
c0

k1

Figure 10. Proposed Mechanical Model of the MR Damper.

F = z + c 0 ( x y) + k 0 ( x y ) + k 1 ( x x 0 )

(10)

From (7), the total force can also be written as


F = c 1 y + k 1 ( x x 0 ) .

(11)

In this model, the accumulator stiffness is represented by k 1 and the viscous damping observed at
larger velocities is represented by c 0 . A dashpot, represented by c 1 , is included in the model to
produce the roll-off that was observed in the experimental data at low velocities, k 0 is present to
control the stiffness at large velocities, and x 0 is the initial displacement of spring k 1 associated
with the nominal damper force due to the accumulator.
The parameters for the model in (9), (10) were chosen to be = 963 N/cm, c 0 = 53
N sec/cm, k 0 = 14 N/cm, c 1 = 930 N sec/cm, k 1 = 5.4 N/cm, = 200 cm2, = 200
cm2, n = 2 , A = 207 , and x 0 = 18.9 cm, which fit the response of the proposed model to the
2.5 Hz data shown in Fig. 3 for the case where the voltage to the current driver was 1.5 V. A comparison between the predicted responses and the corresponding experimental data is provided in
Fig. 11. The proposed model for the damper predicts the behavior of the damper very well in all
regions, including in the region where the acceleration and velocity have opposite signs and the
magnitude of the velocities are small.
In addition to the graphical evidence of the superiority of the proposed model, a quantitative
study of the errors between each of the models and the experimental data. For each of the models
considered here, the error between the predicted force and the measured force has been calculated
as a function of time, displacement and velocity over two complete cycles. The following expressions have been used to represent the errors

E t = -----t- ,
F
Spencer, Dyke, Sain & Carlson

E x = -----x- ,
F
March 10, 1996

E x = -----xF

(12)

11

1500

Force (N)

Experimental
Predicted

1500
0

0.2

0.4

0.6

0.8

12

Time (sec)

Force (N)

a) Force vs. Time


1500

1500

1500
2

1500
12

Displacement (cm)
b) Force vs. Displacement

Velocity (cm/sec)

c) Force vs. Velocity

Figure 11. Comparison Between the Predicted and Experimentally


Obtained Responses for the Proposed Model.

where
T
2
t

( Fexp Fpre ) 2

dt

(13)

0
T
2
x

- dt
( Fexp Fpre) 2 ---dt
dx

(14)

T
2
x

- dt
( Fexp Fpre) 2 ---dt
dx

(15)

T
2
F

( Fexp F )

dt .

(16)

The resulting normalized errors are given in Table 1. In all cases, the error norms calculated for
the proposed model are considerably smaller than those calculated for the other models considered, indicating that the proposed model is superior to the other models for the MR damper considered.

Spencer, Dyke, Sain & Carlson

March 10, 1996

12

Table 1: Error Norms for MR Damper Models.


Model

Et

Ex

E x

Bingham Model

0.154

0.0398

0.133

Gamota and Filisko Model

0.196

0.0717

0.300

Simple Bouc-Wen Model

0.167

0.0585

0.135

Proposed Bouc-Wen Model

0.0351

0.0228

0.0445

Because of its flexibility and numerical tractability for sinusoidal displacement and constant
magnetic fields, this model will be the focus of the remainder of the paper. In the next section, a
generalization will be considered to model the device when the magnetic field and the prescribed
displacements are arbitrary functions of time.

Generalization for Fluctuating Magnetic Fields


All of the data that we have examined previously has been based on the response of the MR
damper when the applied voltage, and hence the magnetic field, was held at a constant level.
However, optimal performance of a control system which utilizes this device is expected to be
achieved when the magnetic field is continuously varied based on the measured response of the
system to which it is attached. To use the damper in this way, a model must be developed which is
capable of predicting the behavior of the MR damper for a fluctuating magnetic field.
To determine a model that is valid for fluctuating magnetic fields, the functional dependence
of the parameters on the applied voltage (or current) must be determined. For instance, the yield
stress of the MR fluid is directly dependent on the magnetic field strength, so the parameter in
(9)(11) is assumed to be a function of the applied voltage. From the experimental results shown
in Fig. 3, the steady state yield level appears to vary linearly with the applied voltage, and have a
nonzero initial value (i.e., at 0 V). This nonzero initial value is due in part to the fluid which by
design has a small yield strength at zero field for stability against gravitational settling, and in part
due to friction in the piston rod seal. The viscous damping constants also vary linearly with the
applied voltage. Therefore, the following relations are proposed
= ( u ) = a + b u , c 1 = c 1 ( u ) = c 1a + c 1b u and c 0 = c 0 ( u ) = c 0a + c 0b u

(17)

where the dynamics involved in the MR fluid reaching rheological equilibrium are accounted for
through the first order filter
u = ( u v )

(18)

and v is the voltage applied to the current driver. Optimal values of a total of fourteen parameters
( c 0a , c 0b , k 0 , c 1a , c 1b , k 1 , x 0 , a , b , , , n , and A ) must be determined for the prototype
MR damper.
A constrained nonlinear optimization was used to obtain these parameters (Spencer, et al.
1996b). The optimization was performed using a sequential quadratic programming algorithm
available in MATLAB (1994). Table 2 provides the optimized parameters for the generalized
Spencer, Dyke, Sain & Carlson

March 10, 1996

13

Table 2: Parameters for the Generalized Model


Parameter

Value

Parameter

Value

c 0a

21.0 N sec/cm

140 N/cm

c 0b

3.50 N sec/cm V

695 N/cm V

k0

46.9 N/cm

363 cm2

c 1a

283 N sec/cm

363 cm2

c 1b

2.95 N sec/cm V

301

k1

5.00 N/cm

x0

14.3 cm

190 sec1

model that were determined to best fit the data in a variety of representative tests, including: 1)
step response, 2) constant voltage/random displacement, and 3) random displacement/random
voltage. In the following paragraphs these tests will be described, and a comparison made between the data and the responses predicted by the proposed model using the parameters given in
Table 2. Simulations were performed in SIMULINK (1994) using the experimentally determined
displacement x and calculated velocity x of the piston-rod in determining the force generated in
the damper model.
The step response tests consisted of applying a triangular displacement to the damper, resulting in regions in which the velocity is nearly constant, and applying a step change in the applied
voltage from 0 to 2.25 V as the damper passes through the middle of the stroke ( x 0 ). The measured displacement and applied voltage are shown in Fig. 12. The measured force is expected to
jump to a new value and remain there when the step in the voltage occurs. Figure 13 compares the
predicted results to the experimental data for the input signals shown in Fig. 12. The model effectively predicts the behavior of the device. The error norms given in Eq. (12) were calculated to be
E t = 0.101 , E x = 0.051 , and E x = 0.107 for this test. This test also verified that the damper
reaches rheological equilibrium within approximately 6 msec after the step voltage is applied.
Note that the sampling rate was increased to 3 kHz in this test to capture the higher frequency
content of the measured responses.
In the second test to verify the model, the damper was excited with a 15 second random displacement record based on an El Centro earthquake acceleration record. A portion of the displacement record of the damper is shown in Fig. 14. The voltage applied to the current driver was a
constant 2.25 V. The sampling rate was set at 2 kHz. The simulated force is compared to the experimental data in Fig. 15. As seen here, the model accurately predicts the behavior of the damper.
For this test, the error norms given in Eq. (12) were determined to be E t = 0.286 , E x = 0.118 ,
and E x = 0.260 .
For the final verification test, the inputs to the device were chosen to be characteristic of the
operating conditions the MR damper will experience when it is applied to a structure in a semi-active control system. Dyke, et al., (1996a,b) have proposed a clipped-optimal control strategy for
controlling a three story model structure with an MR damper. A controlled simulation was performed based on the numerical example in Dyke, et al., (1996a,b). The input control signal and
displacement of the MR damper determined from this simulation are shown in Figure 16. This
Spencer, Dyke, Sain & Carlson

March 10, 1996

14

Displacement (cm)

1
0.5
0
0.5
1
0

0.2

0.4

0.6

0.8

1.2

0.8

1.2

Time (sec)
Applied Voltage (V)

a) Displacement vs. Time


2.5
2
1.5
1
0.5
0
0.5
0

0.2

0.4

0.6

Time (sec)

b) Applied Voltage vs. Time


Figure 12. Inputs Applied to the MR Damper in the Step Response Test.

1500
1000

Force (N)

500
0
500
Experimental
Predicted

1000
1500
0

0.2

0.4

0.6

0.8

1.2

Time (sec)

a) Force vs. Time


1500

1000

1000

500

500

500

500

1000

1000

Force (N)

1500

1500
1

0.5

0.5

Displacement (cm)

b) Force vs. Displacement

1500
10

10

Velocity (cm/sec)

c) Force vs. Velocity

Figure 13. Comparison of Predicted Response and


Experimental Data for Step Response Tests.

Spencer, Dyke, Sain & Carlson

March 10, 1996

15

Displacement (cm)

0.6
0.4
0.2
0
0.2
0.4
0

0.5

1.5

Time (sec)

Figure 14. Displacement Input Applied to MR Damper in


Constant Voltage, Random Displacement Test.

Force (N)

1500

2000
0

Experimental
Predicted
0.5

1.5

Time (sec)

a) Force vs. Time


1500

Force (N)

1500

2000
0.4

0.6

Displacement (cm)

b) Force vs. Displacement

2000
35

25

Velocity (cm/sec)

c) Force vs. Velocity

Figure 15. Comparison of the Model Results and the


Experimental Data for the Constant Voltage, Random
Displacement Tests.

sample displacement and control input history were applied simultaneously to the MR damper. A
comparison between the experimental results and the predicted behavior of the damper is shown
in Figure 17. Again, excellent agreement is found between the experimental and model responses.
For this test, the error norms given in Eq. (12) were determined to be E t = 0.188 , E x = 0.164 ,
and E x = 0.188 .

Spencer, Dyke, Sain & Carlson

March 10, 1996

16

Displacement (cm)

1
0.5
0
0.5
1
0

0.5

1.5

1.5

Time (sec)
Applied Voltage (V)

a) Displacement vs. Time


2.5
2
1.5
1
0.5
0
0.5
0

0.5

Time (sec)

b) Applied Voltage vs. Time

Figure 16. Inputs to MR Damper for the Control Simulation Test.

Conclusion
Recently developed MR fluids have high strength, low viscosity and low power requirements, are stable over a broad temperature range, and are insensitive to impurities commonly introduced during manufacturing. The magnetorheological (MR) damper, consisting of a fixedorifice damper filled with a controllable MR fluid, is an attractive semi-active control device that
appears to have significant potential for structural control applications. To take full advantage of
the unique features of the MR damper, a high fidelity model is needed for control design and analysis.
A review of several idealized mechanical models for controllable fluid dampers has been presented. Subsequently, a new model has been proposed that overcomes a number of the shortcomings of these models and can effectively portray the behavior of a typical magnetorheological
damper. This phenomenological model is based on a BoucWen hysteresis model, which is numerically tractable and is capable of exhibiting a wide variety of hysteretic behaviors. A dashpot
has been added in series with the BoucWen model which creates the nonlinear roll-off observed
in the force as the velocity approaches zero; and an additional spring is incorporated into the model to account for the stiffness of the accumulator present in the prototype MR damper. To obtain a
model that reproduces the behavior of the damper with fluctuating magnetic fields, three parameters are assumed to vary with the applied voltage. Additionally, a first order filter has been incorporated into the model to account for the dynamics involved in the MR fluid reaching rheological
equilibrium. When compared with experimental data, the resulting model was shown to accurateSpencer, Dyke, Sain & Carlson

March 10, 1996

17

Force (N)

2000

Experimental
Predicted

1000
0
1000
2000
0

0.5

1.5

Time (sec)

Force (N)

a) Force vs. Time


2000

2000

1000

1000

1000

1000

2000
1

0.5

0.5

Displacement (cm)

b) Force vs. Displacement

2000
30

30

Velocity (cm/sec)

c) Force vs. Velocity

Figure 17. Comparison of the Model Results and the Experimental


Data for the Control Simulation Test.

ly predict the response of the MR damper over a wide range of operating conditions, including
step voltage, random displacement/constant voltage, and random displacement/random voltage
tests. These results indicate that the model can be effectively used for control algorithm development and system evaluation.

Acknowledgment
This research is partially supported by National Science Foundation Grant Nos. BCS 9301584
and BCS9500301. In addition, the authors from Notre Dame would like to express their appreciation to the Lord Corporation of Cary, North Carolina for providing the prototype magnetorheological damper. In particular, the encouragement and advice of Mr. Thomas Loftus is gratefully
acknowledged.

References
Akbay, Z. and Aktan, H.M. (1990). Intelligent Energy Dissipation Devices. Proc. of the Fourth
U.S. National Conference on Earthquake Engineering, Vol 3, No. 4. pp. 427-435.
Akbay, Z. and Aktan, H.M. (1991). Actively Regulated Friction Slip Devices. Proc. of the 6th
Canadian Conference on Earthquake Engineering, pp. 367-374.

Spencer, Dyke, Sain & Carlson

March 10, 1996

18

Bossis, G. and Lemaire, E. (1991). Yield Stresses in Magnetic Suspensions, Journal of Rheology, Vol 35(7), pp. 13451354.
Carlson, J.D. (1994). The Promise of Controllable Fluids. Proc. of Actuator 94 (H. Borgmann
and K. Lenz, Eds.), AXON Technologie Consult GmbH, pp. 266270.
Carlson, J.D. and Weiss, K.D. (1994). A Growing Attraction to Magnetic Fluids, Machine Design, August, pp. 6164.
Carlson, J.D., Catanzarite, D.M. and St. Clair, K.A. (1995). Commercial Magneto-Rheological
Fluid Devices, Proceedings of the 5th International Conference on ER Fluids, MR Fluids and
Associated Technology, U. Sheffield, UK.
Carlson, J.D. and Chrzan, M.J., U.S. Patent #5,277,281, Magnetorheological Fluid Dampers,
1994.
Carlson, J.D. Chrzan, M.J., and James, F.O., U.S. Patent #5,398,917, Magnetorheological Fluid
Devices, 1995.
Cherry, S. (1994). Research on Friction Damping at the University of British Columbia. Proceedings of the International Workshop on Structural Control, USC Publication Number CE9311, pp. 84-91.
Constantinou, M.C., Symans, M.D., Tsopelas, P. and Taylor, D.P. (1993). Fluid Viscous Dampers
in Application of Seismic Energy Dissipation and Seismic Isolation. Proc. of ATC-17-1 Seminar
of Seismic Isolation, Passive Energy Dissipation, and Active Control, Vol. 2, pp. 581-591.
Constantinou, M.C. and Symans, M.D., (1994). Semi-Active Fluid Viscous Dampers for Seismic
Response Control. Proc. of the First World Conference on Structural Control, Pasadena, CA.
Demchuk, S.A. (1993). Heat Transfer in Narrow Gaps Filled with Magnetorheological Suspensions. Journal of Magnetism and Magnetic Materials, Vol. 122, No. 1 / 3, p. 312.
Dowdell, D.J and Cherry, S. (1994). Semi-active Friction Dampers for Seismic Response Control of Structures. Proc. of the Fifth US National Conference on Earthquake Engineering, Vol. 1,
pp. 819-828.
Dyke, S.J., B.F. Spencer, Jr., M.K. Sain, and Carlson, J.D. (1996a). A New Semi-Active Control
Device for Seismic Response Reduction, Proc. 11th ASCE Engrg. Mech. Spec. Conf., Ft. Lauderdale, Florida.
Dyke, S.J., B.F. Spencer, Jr., M.K. Sain, and Carlson, J.D. (1996b). Seismic Response Reduction
Using Magnetorheological Dampers. Proc. of the IFAC World Congress, San Francisco, California.
Ehrgott, R.C. and Masri, S.F. (1992). Modelling of Oscillatory Dynamic Behavior of Electrorheological Materials in Shear. Smart Materials and Structures, Vol. 4, pp. 275285.
Ehrgott, R.C. and Masri, S.F. (1994). Structural Control Applications of an Electrorheological
Device. Proceedings of the International Workshop on Structural Control, USC Publication
Number CE-9311, pp. 115-129.

Spencer, Dyke, Sain & Carlson

March 10, 1996

19

Fedorov, V. A. (1992). Features of Experimental Research into the Characteristics of Magnetorheological and Electrorheological Shock Absorbers on Special Test Stands. Magnetohydrodynamics, Vol. 28, No. 1, p. 96.
Feng, Q. and Shinozuka, M. (1990). Use of a Variable Damper for Hybrid Control of Bridge Response under Earthquake. Proc. of the U.S. National Workshop on Structural Control Research.
USC Publication No. CE-9013, pp. 107112.
Gamota, D.R. and Filisko, F.E. (1991). Dynamic Mechanical Studies of Electrorheological Materials: Moderate Frequencies. Journal of Rheology, Vol. 35, pp. 399425.
Gavin, H.P., Ortiz, D.S. and Hanson, R.D. (1994). Testing and Modeling of a Prototype ER
Damper for Seismic Structural Response Control. Proceedings of the International Workshop on
Structural Control, USC Publication Number CE-9311, pp. 166-180.
Gavin, H.P., Hose, Y.D., and Hanson, R.D. (1994). Design and Control of Electrorheological
Dampers. Proc. of the First World Conference on Structural Control, Pasadena, CA, August 3-5,
Vol. 1, pp. WP3-83 through WP3-92.
Gavin, H.P. (1994). Electrorheological Dampers for Structural Vibrations Suppression. Ph.D.
Dissertation, The University of Michigan, Department of Civil and Environmental Engineering.
Kamath, G.M., and Wereley, N.M. (1996). A Nonlinear Viscoelastic-Plastic Model for Electrorheological Fluids. Smart Materials and Structures, (submitted).
Grasselli, Y., Bossis, G. and Lemaire, E. (1993). Field-Induced Structure in Magnetorheological
Suspensions. Progress in Colloid & Polymer Science, Vol. 93, p. 175.
Horvath, D. and Kopcansky, P. (1993). Magnetic Dimer Motion Effects in Rotating Magnetic
Field (A Qualitative Model of Magnetoviscosity and Permittivity in Magnetorheological Suspensions). Czechoslovak Journal of Physics, Vol. 43, No. 6, p. 671.
Housner, G.W. and Masri, S.F. (Eds.). (1990). Proceedings of the U.S. National Workshop on
Structural Control Research, USC Publications No. M9013, University of Southern California.
Housner, G.W. and Masri, S.F. (Eds.). (1993) Proceedings of the International Workshop on
Structural Control, USC Publication No. CE9311, Univ. of Southern California.
Housner, G.W., Masri, S.F. and Chassiakos, A.G. (Eds.), (1994). Proceedings of the First World
Conference on Structural Control, Pasadena, CA, August 3-5.
Inaudi, J.A. and Kelly, J.M. (1994) "Experiments on Tuned Mass Dampers Using Viscoelastic,
Frictional and Shape-Memory Alloy Materials. Proceedings of the First World Conference on
Structural Control, Pasadena, CA, August 3-5, Vol. 2, pp. TP3-107.
Kabakov, A.M. and Pabat, A.I. (1990). Development and Investigation of Control Systems of
Magnetorheological Dampers. Soviet Electrical Engineering, Vol. 61, No. 4, p. 55.
Kashevskii, B.E. (1990). Relaxation of Viscous Stresses in Magnetorheological Suspensions.
Magnetohydrodynamics, Vol. 26, No. 2, p. 140.

Spencer, Dyke, Sain & Carlson

March 10, 1996

20

Kawashima, K., Unjoh, S. and Shimizu, K. (1992). Experiments on Dynamics Characteristics of


Variable Damper. Proc. of the Japan National Symposium on Structural Response Control, Tokyo, Japan, p.121.
Kobori, T., Takahashi, M., Nasu, T., Niwa, N. and Ogasawara, K. (1993). Seismic Response
Controlled Structure with Active Variable Stiffness System. Earthquake Engineering and Structural Dynamics, Vol. 22, pp. 925941.
Kordonsky, W.I. (1993a). Elements and Devices Based on Magnetorheological Effect. Journal
of Intelligent Material Systems and Structures, Vol. 4, No. 1, p. 65.
Kordonsky, W.I. (1993b). Magnetorheological Effect as a Base of New Devices and Technologies. Journal of Magnetism and Magnetic Materials, Vol. 122, No. 1 / 3, p. 395.
Kordonsky, W.I. Shulman, Z.P., Gorodkin, S.R., Demchuk, S.A., Prokhorov, I.V., Zaltsgendler,
E.A. and Khusid, B.M. (1990). Physical Properties of Magnetizable Structure-Reversible Media. Journal of Magnetism and Magnetic Materials, Vol. 85, pp. 114120.
Kordonsky, W.I., Gorodkin, S.P. and Demchuk, S.A. (1993). Magnetorheological Control of
Heat Transfer. International Journal of Heat and Mass Transfer, Vol. 36, No. 11, pp. 2783.
Kurata, N., Kobori, T., Takahashi, M. Niwa, N. and Kurino, H. (1994). Shaking Table Experiements of Active Variable Damping System. Proceedings of the First World Conference on Structural Control, Pasadena, CA, August 3-5, pp. TP2-108 through TP2-107.
Lemaire, E., Grasselli, Y. and Bossis, G. (1994). Field Induced Structure in Magneto and Electro-Rheological Fluids. Journal de Physique, Vol. 2, No. 3, p. 359.
Makris, N., Hill, D., Burton, S. and Jordan, M. (1995). Electrorheological Fluid Dampers for
Seismic Protection of Structures. Proc. SPIE Conf. on Smart Struct. and Materials (I. Chopra,
Ed.), San Diego, California, 184194.
Makris, N., Burton, S.A., Hill, D. and Jordan, M. (1996). Analysis and Design of an Electrorheological Damper for Seismic Protection of Structures, Journal of Engineering Mechanics, ASCE,
(in press).
MATLAB (1994). The Math Works, Inc. Natick, Massachusetts.
McClamroch, N.H. and Gavin, H.P. (1995). Closed Loop Structural Control Using Electrorheological Dampers. Proceedings of the American Control Conference, Seattle, Washington, pp.
41734177.
Minagawa, K., Watanabe, T. and Munakata, M. (1994). A Novel Apparatus for Rheological
Measurements of Electro- Magneto-Rheological Fluids. Journal of Non-Newtonian Fluid Mechanics, Vol. 52, No. 1, p. 59.
Mizuno, T., Kobori, T., Hirai, J., Matsunaga, Y. and Niwa, N. (1992). Development of Adjustable Hydraulic Dampers for Seismic Response Control of Large Structure. ASME PVP Conference, PVP-Vol. 229, pp. 163170.
Pabat, A.I. (1990). Controlled Magnetorheological Shock Absorbers. Magnetohydrodynamics.,
Vol. 26, No. 2, p. 222.

Spencer, Dyke, Sain & Carlson

March 10, 1996

21

Patten, W.N., Kuo, C.C., He, Q., Liu, L., and Sack, R.L. (1994). Suppression of Vehicle-Induced
Bridge Vibration via Hydraulic Semiactive Vibrartion Dampers. Proc. of the First World Conference on Structural Control, Pasadena, CA, August 35, pp. FA1-30 through FA1-38.
Rabinow, J. (1948). The Magnetic Fluid Clutch. AIEE Transactions, Vol. 67, pp. 13081315.
Sack, R.L. and Patten, W., (1994). Semiactive Hydraulic Structural Control. Proceedings of the
International Workshop on Structural Control, USC Publication Number CE-9311, pp. 417-431.
Sack, R.L., Kuo, C.C., Wu, H.C., Liu, L. and Patten, W.N. (1994). Seismic Motion Control via
Semiactive Hydraulic Actuators. Proc. of the U.S. Fifth National Conference on Earthquake Engineering, Chicago, Illinois, Vol. 2, pp. 311320.
Savost'yanov, A. (1992). Effects of Magnetomechanical Relaxation in a Magnetorheological
Suspension. Magnetohydrodynamics, Vol. 28, No. 1, p. 42.
Shames, I.H. and Cozzarelli, F.A. (1992). Elastic and Inelastic Stress Analysis. Prentice Hall, Englewood Cliffs, New Jersey, pp. 120-122.
Shinozuka, M., Constantinou, M.C. and Ghanem, R. (1992). Passive and Active Fluid Dampers
in Structural applications, U.S./China/Japan Workshop on Structural Control, Shanghai, China,
pp. 507516.
Shulman, Z.P., Kordonsky, W.I. and Zaitsgendler. (1986). Structure, Physical Properties and Dynamics of Magnetorheological Suspensions. International Journal of Multiphase Flow, Vol. 12,
No. 6, pp. 935955.
Shulman, Z.P., Kordonsky, W.I. and Gorodkin, S.R. (1989). A Recuperator with a Magnetorheological Coolant. Journal of Engineering Physics, Vol. 56, No. 4, p. 438.
SIMULINK (1994). The Math Works, Inc. Natick, Massachusetts.
Spencer Jr., B.F., Dyke, S.J., Sain, M.K. and Carlson, J.D. (1996a). Idealized Model of a Magnetorheological Damper, Proceedings of the 12th Conference on Analysis and Computation, ASCE,
Chicago, Illinois.
Spencer Jr., B.F., Dyke, S.J., Sain, M.K. and Carlson, J.D. (1996b). Nonlinear Identification of
Semi-Active Control Devices. 11th ASCE Engrg. Mech. Spec. Conf., Ft. Lauderdale, Florida.
Stanway, R. Sproston, J.L. and Stevens, N.G. (1985). Non-linear Identification of an Electrorheological Vibration Damper. IFAC Identification and System Parameter Estimation, pp. 195
200.
Stanway, R,. Sproston, J.L. and Stevens, N.G. (1987). Non-linear Modelling of an Electro-rheological Vibration Damper. J. Electrostatics, Vol. 20, pp. 167184.
Soong, T.T. (1990). Active Structural Control: Theory and Practice, Longman Scientific and
Technical, Essex, England.
Soong, T.T., Masri, S.F. and Housner, G.W. (1991). An Overview of Active Structural Control
under Seismic Loads. Earthquake Spectra, Vol. 7, No. 3, pp. 483505.

Spencer, Dyke, Sain & Carlson

March 10, 1996

22

Wen, Y.K. (1976). Method of Random Vibration of Hysteretic Systems. Journal of Engineering
Mechanics Division, ASCE, Vol. 102, No. EM2, pp. 249263.
Winslow, W.M. (1947). Method and Means for Translating Electrical Impulses Into Mechanical
Force. US Patent No. 2,417,850.
Winslow, W.M. (1949). Induced Fibration of Suspensions. Journal of Applied Physics, Vol. 20,
pp. 11371140.

Spencer, Dyke, Sain & Carlson

March 10, 1996

23

Вам также может понравиться